Skip to content

histogram of time-stamp data

8 messages · e-letter, Jessica Streicher, Rui Barradas

e-letter
# 
Readers,

A data set consists of time-stamp values:

00:00:00
23:11:00
06:22:00

The data set was imported:

timestamps<-read.table("path/to/timestampsvalues")
hist(timestamps)

...error... x must be numeric

Then tried:

plot(timestamps).

How to adjust the graph to create a histogram where the intervals
ranges can be specified, e.g. intervals of 60 minutes?

Thanks in advance.
Rui Barradas
# 
Hello,

Try the following.


timestamps <- as.POSIXct(Sys.Date()) + sample(24*60*60, 1e3, TRUE)

h1 <- cut(timestamps, breaks="hour")
h2 <- cut(timestamps, breaks="15 mins")

op <- par(mfrow=c(1, 2))
hist(as.integer(h1))
hist(as.integer(h2))
par(op)


Hope this helps,

Rui Barradas

Em 16-07-2012 09:47, e-letter escreveu:
e-letter
# 
On 16/07/2012, r-help-request at r-project.org
<r-help-request at r-project.org> wrote:
> ------------------------------
 >
 > Message: 77
 > Date: Mon, 16 Jul 2012 10:48:39 +0100
 > From: Rui Barradas <ruipbarradas at sapo.pt>
 > To: e-letter <inpost at gmail.com>
 > Cc: r-help at r-project.org
 > Subject: Re: [R] histogram of time-stamp data
 > Message-ID: <5003E377.3000604 at sapo.pt>
 > Content-Type: text/plain; charset=ISO-8859-1; format=flowed
 >
 >
 > timestamps <- as.POSIXct(Sys.Date()) + sample(24*60*60, 1e3, TRUE)
 >

 Why is it necessary to apply the sample to the current date? Looking
 at the dataframe, I noticed that values have been changed:

 file 'test.txt':
 12:19:00
 09:30:00
 16:56:00
 01:56:00
 10:44:00
 10:31:00
 02:14:00
 05:05:00
 12:52:00
 21:50:00

 R command terminal input:
 > timestamps<-read.table("test.txt")
 > timestamps
          V1
 1  12:19:00
 2  09:30:00
 3  16:56:00
 4  01:56:00
 5  10:44:00
 6  10:31:00
 7  02:14:00
 8  05:05:00
 9  12:52:00
 10 21:50:00
 > timestamps <- as.POSIXct(Sys.Date()) + sample(24*60*60, 1e3, TRUE)
 > write.csv(timestamps,file="test1.txt")

 test1.txt:
 "","x"
 "1",2012-07-16 04:52:48
 "2",2012-07-16 21:21:28
 "3",2012-07-16 18:58:27
 "4",2012-07-16 22:17:25
 "5",2012-07-16 11:13:52
 "6",2012-07-16 03:17:35
 "7",2012-07-16 02:14:17
 "8",2012-07-16 14:18:27
 "9",2012-07-16 14:39:16

 Why is this happening?
Rui Barradas
# 
Hello,

Em 16-07-2012 22:45, e-letter escreveu:
No! That instruction is just to create a data example with more 
date/time values, in this case with a total of 1e3 different values. 
What's important is the way to plot the histogram, namely, the cut() 
with two example time periods, and that hist() needs numbers, not cut's 
levels.

Rui Barradas
e-letter
#
On 16/07/2012, Rui Barradas <ruipbarradas at sapo.pt> wrote:
With the original data provided, R reports an error, that "'x' must be numeric'.
Rui Barradas
# 
Hello,

That's not the error I've had. You must be aware that read.table creates 
a data.frame and therefore the object 'timestamps' is NOT holding time 
stamps, it's holding a vector, 'V1', of time stamps.


timestamps <- read.table(text="
12:19:00
09:30:00
16:56:00
01:56:00
10:44:00
10:31:00
02:14:00
05:05:00
12:52:00
21:50:00
", stringsAsFactors=FALSE)
str(timestamps)

timestamps <- as.POSIXct(timestamps$V1, format="%H:%M:%S")  # here

h1 <- cut(timestamps, breaks="hour")
h2 <- cut(timestamps, breaks="15 mins")

op <- par(mfrow=c(1, 2))
hist(as.integer(h1))
hist(as.integer(h2))
par(op)


And the rest works.

Rui Barradas

Em 17-07-2012 07:11, e-letter escreveu:
e-letter
#
On 17/07/2012, Rui Barradas <ruipbarradas at sapo.pt> wrote:
Was not aware of the significance of the data frame of vector values.
Can confirm the success of adding this parameter to the command 'read.table'
Confirmed, thanks.